Abstract: The spectra fusion method for multi-sections sinusoids（called “the fore-fusion method” for short）is an effective way of estimation frequency for the sinusoids in low Signal-to-Noise Ratio（SNR）, which has an important theoretical significance and practical value. In order to meet the high real-time demand in some fields such as radar, sonar and electronic countermeasures, a fast spectra fusion algorithm for multi-section sinusoids is put forward. This proposed algorithm can reduce the calculation and improve the real-time characteristic by the following techniques：design a fast DTFT algorithm, reduce dimensions of the weighted fusion spectrum matrix, and analyse the correlation of the 1/3 main-lodes of the optimization weighted-accumulation spectrum and the accumulation spectrum. The principles of the above techniques are expatiated. Calculation analyses and simulations demonstrate that the proposed algorithm can reduce most calculation of the fore-fusion method with lower little precision, and it works better in very low SNR[(SNR-13 dB)].

Key words: fast spectra fusion, multi-sections sinusoids, fast Discrete Time Fourier Transform（DTFT） algorithm, dimensions reduction, correlation of 1/3 main-lodes

摘要： 多段正弦信号频谱融合法（简称“原融合算法”）是提高低信噪比条件下正弦信号频率估计精度的一条有效途径，具有重要研究意义和应用价值。为满足雷达、声纳、电子对抗等实时性要求较高的频率估计应用需求，提出多段正弦信号快速频谱融合算法。该方法通过设计离散时间傅里叶变换（Discrete Time Fourier Transform，DTFT）快速算法、降维处理加权融合频谱矩阵和1/3主瓣相关性分析处理等措施来降低算法计算量，提高实时性。重点对上述三项措施的原理进行了阐述与分析。计算量对比和仿真实验表明，多段正弦信号快速频谱融合算法在精度损失极小的前提下，能够大幅降低计算量；在信噪比极低的情况下[(SNR-13 dB)]，其性能略优于原融合算法。

关键词: 快速频谱融合, 多段正弦信号, 离散时间傅里叶变换（DTFT）快速算法, 降维, 1/3主瓣相关